This research aims to uncover the functional and neural processes that underlie reference comprehension, especially with respect to the relation between reference form and referent salience. Explaining how reference is processed is critical for understanding natural language use, which crucially depends on the matching between referential expressions and real world referents. The specific hypothesis tested in this research is that the choice and processing of referential form reflect general memory processes that are involved in the representation of multiple referents and that are prone to interference, especially when the referents are salient. According to this "interference view", the use of general referential forms such as pronouns can reduce this interference and is therefore preferred when referents could be easily identified. At the neural level this hypothesis predicts that repeated full reference to salient referents would increase brain activation both in areas that are involved in the semantic representation of referents (e.g., regions in the temporal lobe for many types of object referents) and in areas that are known to support the manipulation and integration of multiple representations such as the areas around the intra parietal sulcus (IPS). Self paced reading and matching functional MRI (fMRI) experiments will test the central predictions of this view. The first prediction that will be tested is that behavioral and neural interference is associated with repeated reference to a salient referent but not to a non salient referent. The second prediction that will be tested is that not only repeated names (e.g., 'Joe') but also repeated definite description referents (e.g., 'the man') lead to behavioral and neural interference when the referent is salient. The third and last prediction that will be tested is that non repeated category definite description references (e.g., 'the bird'as a reference to 'the robin') lead to reduced interference similar to pronouns. Because little is currently known about the brain basis of referential processing, any data about these issues would be valuable. The application of converging behavioral and fMRI methodologies to this new theoretical account is unique in discourse research and holds promise for bridging functional accounts of discourse processes with their possible brain basis. Therefore, tying referential processing to underlying brain mechanisms would represent a major leap forward in terms of the scientific understanding of these issues. Such understanding is important for the diagnosis and treatment of language disorders, especially ones that involve deficits in referential processing such as Alzheimer's disease. The better understanding of these processes may also lead to methods and strategies for improving life quality in patients suffering from language disorders and their families. PUBLIC HEALTH RELEVANCE Explaining how reference is processed is critical for understanding natural language use, which crucially depends on the matching between referential expressions and real world referents. Tying referential processing to underlying brain mechanisms would represent a major leap forward in terms of the scientific understanding of these issues. Such understanding is important for the diagnosis and treatment of language disorders, especially ones that involve deficits in referential processing such as Alzheimer's disease and may suggest methods and strategies for improving life quality in patients suffering from language disorders and their families.